Fifth Generation

If you have more than just a few snapshots in your photo archive, you need powerful software to manage them. The idea is not only to make the images accessible by filename, but in many other ways, as well, such as by time of recording, by keyword, or by geographic location.

For years, digiKam [1] has been considered the most important free software for managing and manipulate large volumes of images. The freshly published digiKam 5 ditches a fair amount of ballast and includes a number of conceptual changes, as well as many smaller additions. Version 5 largely replaces KDE dependencies with dependencies on Qt5 and introduces a number of new features. Version 5.1, the version used for this article, fixed a number of bugs; versions 5.2-5.4 made several improvements to the similarity search engine and include a complete re-write of video file support. Also new in v5.4 is the Maintenance tool (which runs processes in the background to maintain image collections and database contents) in the Tools menu. In this article, I'll take a detailed look at the fifth generation of digiKam.

The Old and the New

DigiKam is based on a simple, clear-cut concept that is reflected in the design. The main window is divided into three panels (Figure 1): the image source (e.g., an album), the images it contains, and the details of the image(s) selected in the current folder or album. This structure has proven its value.

Figure 1: Structured design: On the left, select the source of the images that appear at the center; the details are shown on the right.

Along the left edge, you choose between individual albums (i.e., folders in the filesystem), keywords, (Exif) data, or search results. Whatever you enable by clicking on the appropriate button in the left bar is displayed by digiKam in the central main window. A second mouse click on the same button hides the left bar and thus expands the main window.

The new version of digiKam adds some new buttons in the left sidebar, which I will look at in more detail later. Showing the number of images in an album is a new function enabled by clicking Settings | Configure digiKam, choosing Album View in the Configure dialog, and checking the Show a count of items in Tree Views entry.

In the main window, digiKam features a new Table preview mode in which all the key data for the images are clearly summarized. Initially, you see a couple of columns, but you can add more with the contextual menu. Now, you also can select a theme for the entire program window in Settings | Themes.

The OpenGL extensions viewer is missing from the Tools menu. To make up for this, the function for creating panoramic views by merging multiple images now works again, although the Setup dialog complains about missing components. For this function to work, you need to install the Hugin program  [2] with the package manager. DigiKam also provides an option for blending images of series of exposures.

DigiKam's main window has hardly changed. To select an image to be processed, either click on individual pictures with the mouse, Ctrl-click to select multiple images, or Shift-click to select whole areas. At top left are the plus and minus buttons for adding images to or removing them from the selection.

On the right side, the familiar tab applies special functions (e.g., adding keywords, accessing Exif tags, etc.) to the selected images. In the new release, version info for images, geodata, a filter to restrict the image selection in the main window, and access to the file attributes were added.

The functions in the status bar at the bottom of the window have also changed: You can no longer specify filters here; instead, the bar only shows the currently active filters, which you can disable by unchecking.

Details

The importance of searching grows as the number of images managed by digiKam increases; thus, you will want to familiarize yourself at an early stage with the buttons in the left sidebar so you can find your way around later when you have a larger image collection.

A search by labels and ratings is intuitive (Figure 2). In case of ratings, simply assign the number of asterisks desired to the images. Because this solution proved to be far too rigid, two new variants were added: picks (denoted by flags) and colors. For example, the flags can indicate the quality of the images (see the "Quality Management" box): The Rejected Item option lets you tag bad images that you only keep in the database because you lack a better image of the subject.

Figure 2: In digiKam 5, images can be tagged with a number of abstract ratings that are easy to filter.

Quality Management

digiKam uses the new flags in a very interesting way: Version 5 has an analysis tool to automatically determine image quality. DigiKam detects data on image sharpness, noise, exposure errors, and compression and formulates a characteristic value. First, enable sorting by image quality and define how exactly to sort under Settings | Configure digiKam  | Image Quality Sorter (Figure 3). Then, choose Tools | Maintenance | Image Quality Sorter to assign quality labels to images according to your settings. Generally, quality analysis can be used only if no other maintenance task (e.g., face recognition) is active. Examining the images takes a huge amount of computing power and plenty of time, in particular for larger image stocks. Also, the analysis blocks the use other digiKam functions.

Figure 3: digiKam automatically determines image quality on request and retrieves the best images from the database without further intervention.

You can assign these flags in different ways: Either select the corresponding flag from the context menu of the main window below Assign Labels or – and this is far faster – use keyboard shortcuts: Alt+1 means "rejected," Alt+2 is "accepted," Alt+3 is "pending," and Alt+0 means you have assigned no rating. Unfortunately, many window managers intercept keyboard shortcuts (e.g., Alt+2 often opens a window in which you can enter a program name). That said, the keyboard shortcuts can be changed easily in Settings | Configure Shortcuts.

Color markings go a step further in the direction of abstraction: You can decide what they mean. For example, you could use them to tag particularly important images. Ten colors can be selected using the keyboard shortcuts Ctrl+Alt+1 through Ctrl+Alt+0. Alternatively, you will also find all these labels in Filters tab in the right sidebar.

Geodata

Generally speaking, you have several ways to add geodata to images. The most universal and reliable method is still to do it manually, but modern cameras usually support automatic geotagging, so you can locate where you captured an image. Even GPS-enabled cameras, though, can make the occasionally error (e.g., failing to find satellites while taking the picture).

To modify the geodata, choose the pictures to be geolocated and call Image | Geolocation | Edit Coordinates. You can edit several images from different points of view in a single step. DigiKam always starts the wizard in overview mode, so you need to take a number of steps to assign the right location (Figure 4). Here, I'm using OpenStreetMap instead of the default, Marble.

Figure 4: If your camera does not add GPS location data automatically to the snapshots, you can do so retroactively in digiKam.

Next, select the images for a location from the list and drag them onto the map at the location you took them. That is typically all you need to do to add the location coordinates. For digiKam to write the positions to the images, go to Settings | Metadata and check the If possible write Metadata to RAW files item. DigiKam now indicates that geolocations are present in the upper right corner of an image in the main window.

To use the Map Searches tab in the left sidebar, Ctrl-select the desired area in the map. Below Search by area: you can either select a region or a specific location (Figure 5). After clicking on image icons in the map, you will find your way around the view quickly.

Figure 5: Searching for images in a given area.

In the map, you can exclude certain areas from the results (e.g., to remove unwanted images from the selection filter). The big Show Non-Geolocated items button lets you track down entries that lack geodata in the database.